Brazil's Electoral Court Sets New Rules to Curb AI Deepfakes in 2026 Vote

Overview

• The TSE approved the resolution on May 20, 2026, banning candidate deepfakes from 72 hours before voting until 24 hours after polls close to prevent last‑minute manipulations.
• The measure makes digital platforms jointly liable if they fail to remove illicit AI content promptly, opening the door to fines or other sanctions for slow takedowns.
• All political ads or materials that use AI‑manipulated image or voice must carry a clear, prominent label and state which technology created the content.
• The resolution also bars AI providers from recommending candidacies to users, a step meant to stop algorithmic nudges that could shape voter choice.
• Officials warn deepfakes can reach millions within hours and say the court will rely on technical teams and talks with academia and international forums to improve detection and enforcement.